AI chip: what are the top startups now?
In our AI chip market deck, you will find everything you need to understand the market
SUMMARY
AI chip: what are the top startups now? The strongest answer is Cerebras, d-Matrix, MatX, Etched, Positron, Fractile, Ayar Labs, Lightmatter, Celestial AI, Tenstorrent, Rebellions, Axelera AI, SiMa.ai and EnCharge AI, but they matter for very different reasons.
The AI chip startup market is no longer one simple Nvidia-challenger race. The most serious companies now cluster around inference, LLM-specific ASICs, memory movement, optical interconnect, edge AI, sovereign infrastructure and moonshot compute architectures.
Cerebras is the clearest maturity outlier. It has revenue, IPO scrutiny and customer demand, which makes it closer to a public AI infrastructure company than a normal startup still selling a future roadmap.
d-Matrix is probably the hottest production-stage private AI chip startup right now. Its late-2025 $275 million round matters, but the bigger signal is that Corsair entered full production in June 2026.
The market is rewarding narrow specialization more than broad accelerator promises. MatX and Etched stand out because they are betting that LLM workloads are now predictable enough to justify dedicated hardware instead of flexible GPU-like systems.
Inference is becoming less about theoretical speed and more about deployable economics. Positron, Fractile and d-Matrix are interesting because they each attack the cost, latency or memory pain of serving increasingly complex AI outputs.
The memory wall may be the most important hidden theme. A startup can show impressive compute performance and still lose if its system cannot move data fast enough, cheaply enough or efficiently enough.
Optical infrastructure is moving from science-project territory into core AI infrastructure. Ayar Labs, Lightmatter and Celestial AI matter because the bottleneck is shifting from chips alone to the roads between chips, memory and racks.
Edge AI remains real but fragmented. Axelera AI, SiMa.ai, Hailo and EnCharge AI are worth tracking, but the category needs harsher judgment because robots, cameras, factories and client devices do not all want the same chip.
Non-U.S. AI chip startups are no longer symbolic. Rebellions gives Korea a serious sovereign infrastructure champion, while Axelera AI and Fractile show Europe now has credible edge and datacenter inference bets.
Funding momentum is strongest where the bottleneck is most painful. Fresh capital is flowing into LLM-specific ASICs, inference efficiency, optical interconnect and sovereign AI infrastructure rather than generic “AI accelerator” stories.
The cleanest ranking uses proof categories rather than one flat leaderboard. Cerebras leads on business proof, d-Matrix on production proof, Positron on young inference traction, Ayar Labs on optical scale-up, and MatX and Etched on bold LLM-specific ASIC appetite.
This market map, featured in our AI chip market deck, highlights top companies and startups in the AI chip market
Which AI chip startups are already too serious to ignore today?
Cerebras, d-Matrix, Groq, Tenstorrent and Rebellions are the names we have to start with, but they are not hot for the same reason.
Cerebras is the most mature name in the group. It is now closer to a public AI infrastructure company than a classic startup. The reason is simple: we have real business numbers. Its 2025 revenue was reported around $510 million, up about 76% year over year, and its IPO gave the market a live benchmark for how much investors are willing to pay for a real Nvidia alternative. That separates Cerebras from almost everyone else here. Most AI chip startups still ask us to believe in future deployments. Cerebras can point to revenue, public-market scrutiny, and major cloud demand.
d-Matrix is the hotter startup story right now. Cerebras is bigger, but d-Matrix is the one that feels more like a fresh market pulse. In November 2025, it raised $275 million at a $2 billion valuation. Then in June 2026, its Corsair inference platform entered full production, with the company claiming up to 10x faster inference and 5x better energy efficiency than standalone Nvidia GPUs on targeted workloads. We should still treat vendor claims carefully, but production changes the conversation. A startup with a big round is interesting. A startup with a big round and hardware entering volume production is much more serious.
Groq still matters because it made fast LLM inference feel like a category before many others did. Groq is famous, well funded, and technically credible. The issue is that the market has caught up to its core message. Today, d-Matrix and Positron feel more urgent because their recent signals are closer to production, funding acceleration, or system-level deployment.
Tenstorrent remains one of the most interesting platform bets. It is broader than d-Matrix or Positron: RISC-V, chiplets, developer stack, training and inference systems. That gives it a bigger possible surface area, but also more execution risk. A narrow inference startup can win one painful use case faster. Tenstorrent is trying to build a more open alternative compute platform, which is harder but potentially more durable.
Rebellions is the strongest non-U.S. infrastructure challenger in this first group. Its September 2025 $250 million Series C at a $1.4 billion valuation, backed by Arm and Samsung, matters because it is not just financial validation. It shows Korea is trying to build a serious domestic AI chip champion. Compared with Groq or Cerebras, Rebellions is less globally proven. Compared with most sovereign AI chip projects, it is unusually well capitalized and commercially concrete.
If you want more recent data on this point, please see our latest AI chip market report.
Which AI chip startups are winning the inference race right now?
In the inference race, d-Matrix is first, Positron is the fastest riser, Fractile is the most exciting pre-commercial bet, and Groq still owns a lot of mindshare.
The reason d-Matrix goes first is timing. In AI chips, being “almost ready” is not enough. Startups die in the gap between great architecture and manufacturable systems. d-Matrix crossed an important line in June 2026 when Corsair entered full production. That does not prove it has won the market, but it puts the company ahead of the many inference startups still living mostly on roadmap credibility.
Positron is the one we would watch most closely after d-Matrix. Its funding path moved very fast: $51.6 million Series A in July 2025, then $230 million Series B in February 2026 at a valuation above $1 billion. That kind of jump in seven months says investors are not just buying a nice technical story. They are betting that Positron’s memory-heavy approach can reduce the pain of running LLM inference at scale. Compared with d-Matrix, Positron looks earlier. Compared with Fractile, it looks more commercially immediate.
Fractile is earlier again, but the signal is too big to ignore. In May 2026, it raised $220 million to build inference chips for frontier models. The company’s core argument is very sharp: as models get more useful, users wait longer for answers, and that waiting time becomes an economic bottleneck. That is a better thesis than “we have a faster chip.” It connects directly to product behavior in AI apps. The caveat is equally clear: Fractile is not yet in the same proof category as d-Matrix. It is hot because of the size of the round, the clarity of the bottleneck, and the quality of the investor bet, not because it has already shown broad customer deployment.
Groq still belongs in the top pack because it trained the market to care about tokens per second and low-latency inference. But these days the stronger question is “who can ship enough useful systems into real AI workloads?” On that question, Groq is still relevant, but d-Matrix and Positron currently give us fresher signals to work with.
Rebellions and FuriosaAI are also worth watching, especially in Korea. Rebellions has the stronger capital base and ecosystem backing. FuriosaAI has the louder strategic signal because it reportedly rejected Meta’s acquisition interest and later worked with LG AI Research around domestic AI workloads. In a global ranking, they sit behind the U.S. inference leaders. In a sovereign AI infrastructure ranking, they move up quickly.
As this chart shows, and as featured in our AI chip market deck, search interest in AI chips has grown significantly
Which AI chip startups are making the best narrow ASIC bet?
MatX and Etched are the two clearest bets, with Fractile close behind.
MatX is the cleanest “new serious player” signal in AI chips right now. In February 2026, it raised $500 million in Series B, led by Jane Street and Situational Awareness, to build chips optimized for large language models. That is not normal funding for an early chip startup. It says the market believes LLM-specific hardware may deserve a dedicated compute stack, not just better GPUs.
Etched is the bolder version of the same idea. Its Sohu chip is designed specifically for transformer workloads. The reported January 2026 financing, around $500 million at a $5 billion valuation, is aggressive because Etched is not promising maximum flexibility. The promise here is more about the dominant AI workload being predictable enough to justify a much narrower chip. That is a brave bet. If transformers stay central, Etched could look brilliant. If architectures shift faster than expected, the same focus becomes a weakness.
MatX feels more institutionally credible. Etched feels more radical. That distinction matters. MatX has the former-Google TPU aura and a large manufacturing-scale round. Etched has the sharper “we are deleting GPU flexibility” thesis. Both are hot, but for slightly different reasons.
Fractile also belongs here, although it is more about inference latency and memory movement than pure transformer-only specialization. It is interesting because it does not sound like another generic accelerator company. It is making a strong claim about where AI products are breaking today: not model quality, but the time and cost of serving increasingly complex outputs.
The takeaway is probably that the market is rewarding focus.
Five years ago, a startup wanted to say its chip could run everything. Today, the most exciting new ASIC startups are saying almost the opposite: we know the workload, we know where the GPU wastes energy, and we are willing to specialize hard.
If you want more recent data on this point, please see our latest AI chip market report.
Which AI chip startups are closest to solving the memory and data-movement bottleneck?
d-Matrix, Positron and Fractile lead on inference memory efficiency, while Ayar Labs, Lightmatter and Celestial AI lead on moving data across AI systems.
This is probably the most important section. The AI chip market is still described as a compute race, but today the pain is often memory, bandwidth, interconnect and power. A chip can have impressive math performance and still lose if it cannot feed the model fast enough.
d-Matrix leads the inference-memory group because its whole architecture is built around digital in-memory compute. The June 2026 production signal matters here because memory-centric chips are easy to admire on paper and hard to productize. d-Matrix has at least moved from thesis to hardware entering volume production.
Positron is different. It is not trying to sound as exotic as some memory-compute startups. Actually, its pitch is more practical: give AI operators a system that can serve LLM inference with better power and cost characteristics. That may be less flashy, but it is exactly how infrastructure buyers think. They do not buy “architecture” but lower cost per useful output.
Fractile is the sharper long-context bet. It is basically saying that as models reason longer, memory becomes the choke point. That is why its May 2026 $220 million round matters. Investors are not just funding another accelerator but a view that inference will become more memory-bound as AI products become more agentic and interactive.
Then there is the optical group. Ayar Labs, Lightmatter and Celestial AI are not trying to beat Nvidia with a single accelerator. Instead, they are attacking the roads between the accelerators. Ayar’s March 2026 $500 million Series E at a $3.75 billion valuation is the strongest fresh signal in that group. Lightmatter remains the higher-profile photonics platform, especially around optical interconnect for data centers. Celestial AI has one of the clearest “memory fabric” stories.
This chart, featured in our AI chip market deck, shows annual VC investment in AI chip startups
Which optical AI infrastructure startups look most important now?
Ayar Labs is the freshest optical winner, Lightmatter is the established category leader, and Celestial AI is the memory-connectivity specialist.
Ayar Labs moved up the ranking in March 2026. Its $500 million Series E brought total funding to about $870 million and valued the company at $3.75 billion. More importantly, the company said the money would go toward high-volume production and test capacity for co-packaged optics. That is the key phrase.
Lightmatter is still the brand leader in photonic AI infrastructure. Its valuation has been higher, its Passage interconnect story is well known, and it sits in the part of the stack where hyperscalers are desperate for relief: bandwidth without exploding power use. Compared with Ayar, Lightmatter feels broader and more famous. Compared with Ayar today, it does not have the freshest funding signal.
Celestial AI is the most interesting if we focus specifically on memory connectivity. Its Photonic Fabric thesis is easy to understand: AI systems are becoming limited by how fast and efficiently data can move between memory and compute. That puts Celestial closer to the “memory wall” problem than to generic optical networking.
Enfabrica is worth mentioning, but in a different way. Its AI networking chip became strategically important enough to draw Nvidia interest. That is a market signal more than a normal startup-growth signal. It tells us the incumbents are not waiting politely for interconnect startups to mature. If the technology matters, they may license it, partner with it, or absorb the people behind it.
If you want more recent data on this point, please see our latest AI chip market report.
Which edge AI chip startups are still worth caring about?
Axelera AI is the freshest edge AI name, SiMa.ai is the most focused physical-AI bet, Hailo is the established leader, and EnCharge AI is the technical wildcard.
Edge AI is a tougher category than datacenter inference. The story sounds attractive: run AI locally, save power, reduce latency, protect privacy. But the market is fragmented. A robot, a camera, an industrial machine and a laptop do not all want the same chip. That is why we should be harsher here than in datacenter inference.
Axelera AI currently has the strongest fresh signal. In February 2026, it raised more than $250 million, bringing total funding since 2021 to about $450 million. That is unusually large for a European edge AI chip company. The presence of BlackRock and Samsung Catalyst Fund also makes the round feel more strategic than a normal venture financing.
SiMa.ai is slightly less hot on funding size, but arguably cleaner on use case. Its August 2025 $85 million round brought total funding to $355 million, and its positioning around physical AI is more specific than “edge AI everywhere.” That matters. The strongest edge AI startups today are about environments where cloud inference is awkward: robots, factories, cameras, machines.
Hailo remains credible, especially for edge AI accelerators and local generative AI. But compared with Axelera and SiMa.ai, the recent excitement feels less sharp. Hailo is still a leader, but it is just not the freshest discovery.
EnCharge AI is the technical wildcard. Its analog in-memory compute approach could be powerful for client and edge AI if it works commercially. The February 2025 $100 million Series B is a real signal. But analog AI hardware has a long history of looking beautiful in theory and difficult in deployment. We would watch EnCharge closely without ranking it above companies already clearer on product channel.
This chart, featured in our AI chip market deck, shows how Nvidia is leading in AI chips
Which non-U.S. AI chip startups are becoming serious?
Rebellions, Axelera AI, Fractile and FuriosaAI are the non-U.S. AI chip startups that matter most right now.
Rebellions is the strongest answer. It has the capital, the Samsung and Arm backing, and the national-infrastructure logic. Its September 2025 $250 million Series C at a $1.4 billion valuation makes it much more credible than the average sovereign AI hardware story. Korea wants an AI chip champion that can sit inside domestic infrastructure.
Axelera AI is Europe’s strongest edge AI signal. The February 2026 $250 million-plus round gives it a scale that most European semiconductor startups never reach. It is not trying to beat Nvidia in frontier training, which is probably wise. Its market is edge acceleration, where deployment channels and power efficiency matter more than owning the whole data-center stack.
Fractile is Europe’s most exciting datacenter inference startup. It is earlier than Axelera commercially, but the ambition is bigger. A $220 million Series B for a UK AI inference chip company would have looked unusual a few years ago. Today, it feels like a response to a real gap: Europe wants more than regulation and model labs. It wants infrastructure leverage.
FuriosaAI is smaller than Rebellions, but the Meta acquisition-interest story made it stand out. A startup does not reject that kind of reported offer unless it believes the independent path can be larger. The LG AI Research partnership gives the company a more concrete domestic route.
The U.S. still has the deepest AI chip startup bench. But today, the non-U.S. list is no longer symbolic. Korea has real AI chip ambition. Europe has serious edge and inference bets. The gap is still large, but it is not empty.
If you want more recent data on this point, please see our latest AI chip market report.
Which AI chip startups are still moonshots, but real enough to track?
Extropic, EnCharge AI, Neurophos, Euclyd and Fractile are the names to watch, with very different risk levels.
Extropic is the most radical. It is trying to build thermodynamic computing for probabilistic AI. This is the kind of company we should not over-rank, because the commercialization gap could be huge. But it deserves attention because it asks a useful question: if AI becomes more probabilistic and energy-limited, do we need a very different machine?
EnCharge AI is a more practical moonshot. Analog in-memory compute is still hard, but the market entry point makes sense: client and edge AI, where energy efficiency can matter more than raw datacenter scale. Its 2025 Series B gives it enough capital to keep proving the thesis.
Neurophos and Euclyd are more speculative. They have exciting technical stories around optical or memory-centric compute, but the external proof is still thin. This is where we should avoid pretending confidence. A technical claim is not the same as a market signal.
Fractile is the unusual case. It sits between moonshot and emerging leader. The company is pre-commercial, so we cannot rank it like d-Matrix. But its round size, investor base and very direct inference-latency thesis make it much more than a science project.
We believe these companies could matter a lot, but the ranking should stay below startups with production, revenue, or clear customer pull.
This chart, featured in our AI chip market deck, shows annual funding in AI chip startups
Which AI chip startups have the strongest funding momentum today?
MatX, Ayar Labs, Positron, Fractile, Axelera AI, d-Matrix, Etched and Rebellions have the strongest funding momentum in AI chips right now.
MatX and Etched sit at the top if we only look at narrow LLM-chip appetite. MatX raised $500 million in February 2026. Etched reportedly raised around $500 million at a $5 billion valuation in January 2026. The comparison is useful: MatX looks like the more institutional TPU-style bet; Etched looks like the more aggressive transformer-only bet. Both show that investors are willing to finance specialization when the workload is large enough.
Ayar Labs leads the interconnect side. Its March 2026 $500 million Series E is just as important as the MatX and Etched rounds, but for a different reason. It says the market is now funding optical I/O as core AI infrastructure, not a side technology.
Positron and Fractile lead the next wave of inference startups. Positron’s $230 million Series B came with a clearer near-term system story. Fractile’s $220 million Series B came with a bigger architectural promise but less commercial proof. That is the tradeoff: Positron feels closer to buyers today; Fractile feels like the higher-upside technical bet.
Axelera AI leads edge funding after its February 2026 $250 million-plus round. d-Matrix remains one of the strongest because its November 2025 $275 million round is now paired with production progress. Rebellions is the strongest sovereign funding signal after its September 2025 $250 million round with Arm and Samsung.
As seen above, we should not treat funding as proof of winning. But funding does show where the market is willing to underwrite risk. Right now, capital is flowing toward LLM-specific chips, inference economics, optical interconnect and sovereign infrastructure.
Which AI chip startups have the best real-world proof?
Cerebras is first on business proof, d-Matrix is first on production proof, Positron has the best young inference proof, and Ayar Labs has the best optical scale-up proof.
Cerebras has the strongest business evidence because it has revenue, an IPO and large customer demand. That makes it the hardest company to dismiss. The market may debate valuation, customer concentration and long-term competition, but it cannot dismiss Cerebras as vaporware.
d-Matrix has the cleanest recent production proof. A startup saying it will build a chip is normal. A startup saying its platform is in full production in June 2026 is a different level of evidence. The next proof point will be customer usage at scale, but d-Matrix is clearly ahead of pre-silicon inference startups.
Positron has the best early commercial proof among the newer inference names. The funding ramp is strong, and the system positioning is specific enough to compare against Nvidia economics. It is not yet as proven as d-Matrix, but it is more grounded than startups that only talk about future chips.
Ayar Labs has the strongest optical scale-up proof because its latest round is explicitly tied to high-volume production and test capacity. That matters more than a lab demo. Optical interconnect has been “promising” for years. The current question is who can manufacture and deploy it into AI infrastructure.
Tenstorrent, Rebellions, FuriosaAI and Axelera AI all have credible proof, but with more context needed. Tenstorrent is a broad platform bet, so evidence comes through ecosystem and system partnerships. Rebellions has serious national and strategic backing. FuriosaAI has strategic customer interest. Axelera has edge funding and global commercial expansion. None of them should be dismissed, but their proof is less clean than Cerebras revenue, d-Matrix production or Ayar’s scale-up financing.
If you want more recent data on this point, please see our latest AI chip market report.
This chart, featured in our AI chip market deck, compares the main business model options for AI accelerator chip companies
So who are the top AI chip startups now?
Clearly, the top AI chip startups now are Cerebras, d-Matrix, MatX, Etched, Positron, Fractile, Ayar Labs, Lightmatter, Celestial AI, Tenstorrent, Rebellions, Axelera AI, SiMa.ai and EnCharge AI.
If we make the ranking sharper, we get five groups.
Cerebras is the mature leader. It has moved beyond the normal startup evidence problem because we can judge it on revenue, IPO performance and customer demand. It is not the freshest discovery, but it is the strongest proof point that a non-Nvidia AI infrastructure company can become genuinely large.
d-Matrix is the hottest production-stage private AI chip startup right now. The reason is not only its $2 billion valuation but the combination of late-2025 capital and June 2026 production. That pairing makes it more convincing than startups that only have a big round.
MatX and Etched are the hottest LLM-specific ASIC bets. MatX feels more credible as an institutional chip-building effort. Etched feels more extreme and potentially more explosive. Both are betting that AI workloads have become predictable enough to reward specialization.
Positron and Fractile are the two emerging inference names to watch most closely. Positron looks nearer to deployment. Fractile looks more technically ambitious. If we had to pick the safer near-term watch, it is Positron. If we had to pick the more asymmetric one, it is Fractile.
Ayar Labs, Lightmatter and Celestial AI are the top AI infrastructure bottleneck startups. They may not sound as exciting as “new GPU challenger,” but they sit where the pain is moving: bandwidth, optical I/O, memory fabric and power.
Tenstorrent is the open-platform wildcard. Rebellions is the sovereign AI chip leader. Axelera AI and SiMa.ai are the edge AI names with the clearest current relevance. EnCharge AI is the analog-compute moonshot that deserves tracking.
| Category | Startups selected and why |
|---|---|
| Datacenter inference | d-Matrix ranks first because Corsair entered production; Positron is the fastest riser with a $230 million Series B; Fractile is the high-upside pre-commercial bet; Groq remains relevant but less fresh |
| LLM-specific ASICs | MatX looks like the institutional LLM-chip bet; Etched is the sharper transformer-only bet; Fractile adds a memory-latency angle rather than a generic accelerator story |
| Memory and data movement | d-Matrix, Positron and Fractile attack inference memory pain; Ayar Labs, Lightmatter and Celestial AI attack the system-level bottleneck through optical I/O and photonic fabric |
| Optical AI infrastructure | Ayar Labs has the freshest production-scale funding; Lightmatter remains the category leader; Celestial AI is strongest on memory fabric; Enfabrica shows how strategic AI networking IP has become |
| Edge and client AI | Axelera AI has the freshest edge funding signal; SiMa.ai is the focused physical-AI bet; Hailo remains established; EnCharge AI is the technical wildcard |
| Non-U.S. AI chip startups | Rebellions leads Korea; Axelera AI and Fractile lead Europe on edge and inference; FuriosaAI is a smaller but strategically loud Korean name |
| Moonshots | Extropic is the radical thermodynamic bet; EnCharge AI is the practical analog bet; Neurophos and Euclyd are interesting but still need stronger external proof |
| Funding momentum | MatX, Etched, Ayar Labs, Positron, Fractile, Axelera AI, d-Matrix and Rebellions show where fresh capital is flowing: specialization, inference, interconnect and sovereign infrastructure |
| Real-world proof | Cerebras leads on revenue and IPO proof; d-Matrix on production; Positron on young inference-system traction; Ayar Labs on optical scale-up; Tenstorrent and Rebellions remain broader strategic bets |
OUR METHODOLOGY
This analysis tests which AI chip startups are the most serious today based on the evidence available now. We compare startups across inference, LLM-specific ASICs, memory and data movement, optical infrastructure, edge AI, non-U.S. challengers, moonshots, funding momentum, and real-world proof.
We did not treat the AI chip startup landscape as one simple leaderboard. A company with revenue, a company with production hardware, a company with a $500 million funding round, and a company attacking optical interconnect are not proving the same thing.
For each dimension, we looked for concrete signals: production milestones, reported revenue, major funding rounds, valuation changes, strategic backers, customer or ecosystem validation, and the specific infrastructure bottleneck each company is trying to solve.
We gave more weight to signals that reduce execution uncertainty. Revenue, IPO scrutiny, full production, high-volume production plans, strategic customer interest, and commercially specific system positioning count more than broad claims about having a faster chip.
Funding momentum was treated as a market signal, not as proof of winning. A large round matters because it shows where investors are willing to underwrite risk, but it does not prove customer adoption, manufacturing success, or durable technical advantage.
Production proof and business proof were separated on purpose. Cerebras leads on business proof because it has revenue and public-market scrutiny, while d-Matrix leads on recent production proof because Corsair entered full production in June 2026.
We also separated datacenter, edge, optical and sovereign infrastructure startups because their evidence looks different. Edge AI startups need product-channel clarity, optical startups need scale-up and manufacturing proof, and sovereign AI chip startups need ecosystem and strategic backing.
The final ranking is therefore a structured aggregation of recent signals rather than a reputation ranking. It highlights which startups are serious now, why they are serious, and what kind of proof each one still needs.
Key sources used for this analysis include: d-Matrix on Corsair entering full production, d-Matrix on its $275 million Series C, MatX on its Series B, TechCrunch on MatX’s $500 million round, Data Center Dynamics on Etched’s reported financing, Business Wire on Positron’s $230 million Series B, Fractile on its $220 million raise, Ayar Labs on its $500 million Series E, Lightmatter on Passage M1000, Business Wire on Celestial AI’s Photonic Fabric funding, Axelera AI on its $250 million-plus funding, SiMa.ai on its $85 million round, Rebellions on its $250 million Series C, EnCharge AI on its Series B, TechCrunch on FuriosaAI’s reported Meta offer, Tech.az on FuriosaAI’s LG AI Research partnership, Tom’s Hardware on Cerebras’ IPO filing and revenue context, and MarketWatch on Cerebras’ IPO context.
This chart, featured in our AI chip market deck, shows how revenue is split across customer segments in the AI chip market
Related blog posts
- How strong is fundraising in the AI chip market right now?
- Which companies have raised the most funding in the AI chip market?
- Which companies are the most valued in the AI chip market?
