Nvidia (NVDA) stock investors have had a lot to celebrate in 2026, but it hasn’t been a straight line.
Its stock closed out Aug. 19 at $217.56, down 1%, leaving the stock around 17% higher year to date but slightly lower over the past three months, according to Yahoo Finance.
Nevertheless, shares swung from a May record of $236.54 to $190 in late July, as concerns grew about AI valuations, China, and complex AI financing deals testing the bull case.
Now comes the next big test for the AI behemoth.
Nvidia reports fiscal Q2 results on Aug. 26, with Wall Street expecting $92 billion in sales and $2.08 in earnings per share, roughly double year-ago levels.
Michael Burry, meanwhile, remains a recurring headache for AI bulls. The “Big Short” investor has continued to attack Nvidia’s financing agreements and disclosed bearish bets against giants like Oracle (ORCL) and Nebius (NBIS) as part of his broader AI-bubble thesis.
However, his latest concern hits differently.
He points to a fast-growing Nvidia rival, valued at around $21 billion, that Burry believes could pose “serious competition” to the AI-chip giant.
Burry says Nvidia’s moat faces a real test
In a Substack post covered by Business Insider, Burry just pointed to a serious Nvidia competitor that’s posting some insane performance, cost, and execution speed numbers.
The company, Etched, an AI-chip startup, just raised a massive $700 million in a funding round involving Jane Street, taking its reported valuation to $21 billion, as reported by Reuters. Burry pointed to the company after reviewing reports on its hiring and chip deployment, including pertinent information from a company insider.
What caught his eye in particular is how swiftly Etched has moved.
More Manager Buy/Sells:
- Michael Burry increases his bet against popular chip giant
- Warren Buffett reveals he broke his own investing pattern
- Mark Cuban bets on MLB with Athletics minority stake
The startup said it took it just 44 days to get its AI chips running inference workloads, compared with a process that usually takes six months or longer.
Then there is the talent angle.
Around 15% of Etched’s workforce consists of former Nvidia employees, offering the startup’s engineers direct experience inside the company that’s become the poster child of the AI market.
Burry hailed Etched as “serious competition for NVDA” and said a semi-insider told him the startup’s technology showed nearly 10 times the performance at a lower cost per die. Moreover, he added, the speed of Etched’s progress underscores that “disruption is in the pipeline.”
The implication is obvious: If a smaller rival could deliver meaningfully better economics while scaling quickly, Nvidia’s moat comes under immense pressure from more than just the typical large-chip competitors.
Nvidia’s competition is getting stronger
Etched is far from Nvidia’s only competitive threat.
In fact, a more pertinent question for investors is where Nvidia might lose market share, as its rivals usually attract different parts of the AI-compute stack.
Advanced Micro Devices (AMD) is perhaps the most obvious direct competitor.
The tech giant’s new Helios rack-scale platform layers 72 MI455X accelerators with EPYC CPUs and AMD networking, offering a strong alternative to Nvidia’s strategy of selling complete AI systems instead of stand-alone GPUs.
In addition, AMD claims Helios could potentially deliver up to 30% more inference tokens per dollar compared to Nvidia’s Vera Rubin NVL72 on a selected workload. Also, major hyperscalers such as Meta Platforms (META) plan large-scale AMD deployments, while OpenAI expects to bring Helios online in late 2026.
However, AMD’s big weakness is still software.
Nvidia’s CUDA ecosystem, networking, and years of developer adoption make switching incredibly expensive, which is why Bloomberg Intelligence forecasts Nvidia to retain at least 70% of AI training, despite the competition.
Another major threat comes from its own customers.
Google parent Alphabet (GOOGL) and Amazon (AMZN) are currently designing tailor-made chips specifically for their workloads, where specialization could potentially beat the economics of a general-purpose GPU.
Additionally, Google’s TPU 8i reportedly offers 80% better performance per dollar compared to its previous generation, while TPU 8t targets massive-scale training.
Amazon’s Trainium3 follows a similar logic.
AWS says its latest systems could deliver up to 4.4 times Trainium2’s performance and remarkably lower training and inference costs.
Broadcom (AVGO) is a key enabler of that shift, allowing hyperscalers to build purpose-designed accelerators. OpenAI recently showed off its Broadcom-developed inference chip, while Google is expanding custom-chip production with Marvell (MRVL).
Startups add another major dynamic.
In particular, Cerebras is targeting ultra-fast inference with wafer-scale processors and already boasts a major OpenAI relationship, although its lower margins point to how testing it is for companies to scale up as Nvidia has.
What Nvidia investors should watch next
For Nvidia stock investors, the competitive outlook becomes a lot more complex, but it’s far from being a broken thesis.
The bear case is that if competitors such as Etched genuinely deliver better performance at lower costs, while AMD and other hyperscalers improve their chips, Nvidia loses market share across different parts of AI.
The bigger issue is margin-related pressures, and even if Nvidia retains leadership, it could fake weaker pricing power if customers gain more alternatives. Even modest share losses on the back of aggressive pricing matter a lot for a stock that’s trading at 24-times non-GAAP earnings, according to Seeking Alpha.
The bull case is that Nvidia’s moat is far beyond the chips it’s producing. New players need to prove they can reliably manufacture at scale, secure supply, develop software ecosystems, and support larger customers.
In this scenario, Nvidia will continue racking up comfortable top- and bottom-line beats while rivals remain niche, and the competitive threat remains mostly theoretical. However, if the competition gets more serious, we could see Nvidia’s valuation multiples compress over time.
The worst-case scenario would be if there’s sluggishness in revenue growth, weaker gross margins, and dropping customer dependence on Nvidia.
For now, investors need to monitor competitive traction closely while continuing to separate execution from headline-level disruption.
Related: Bank of America’s latest Nvidia alert is a must-read for worried investors
