Baird Reiterates Outperform on NVIDIA, $500 Price Target

Baird analyst Tristan Gerra reiterated an Outperform rating and $500 price target on NVIDIA (NASDAQ: NVDA) on September 1, 2026.

“WSJ reports that Anthropic entered into a $35B cloud compute agreement with Nvidia-backed neocloud provider Lambda.

Hut 8 (HUT, not covered) will develop the data center, after signing an agreement with Nvidia a few weeks ago to secure capacity. Hut 8 operates the Beacon Point data center leases in Texas.

It is reported that Beacon Point leases have a $19.6B contractual value, and a $50.2B value assuming renewal options are exercised.

Lambda operates 14 data centers (to our count) offering Nvidia-GPU powered AI factories.

It was reported three days ago that Lambda raised about $1B in private short-term debt for computing chips tied its collaboration with Microsoft involving Microsoft leasing ten of thousand of NVDA GPUs.

In 2025, Nvidia signed a $1.5B contract with Lambda to lease back 18k GPU-based servers previously leased to Lambda.

Lambda is also in discussions to raise $3B in pre-IPO money to repay an estimated $1.65B of private financing.

Broadcom has a stated 1GW engagement with Anthropic for 2026, with an April agreement for an additional 5GW capacity engagement starting in 2027.

Broadcom has also announced a partnership with investors including Apollo and Blackstone for the deployment of over 20GW of compute capacity through 2028 with the first tranche of this platform valued at $35B, with Broadcom reportedly raising $70-$80B in debt to support AI frontier labs including Anthropic.

Net, Anthropic is engaged in a $35B cloud compute agreement within the Nvidia AI ecosystem.

Broadcom has announced a 6GW agreement with Anthropic along with over 20GW of compute capacity part of which will be dedicated to Anthropic.

Microsoft has committed $5B into Anthropic, AWS has committed $8B (Anthropic uses both AWS and Trainium), and Google $2-$3B (Anthropic uses 1M+ Google TPUs).

It is difficult for us to tell the mix of GPUs versus TPUs and other custom ASICs used by Anthropic going forward, but we expect Nvidia to retain a compute lead (both training and inferencing) at Anthropic for years to come, with the above-numbers leading us to estimate Nvidia has secured about double the GW capacity at Anthropic versus other AI architecture suppliers, both via Lambda along with agreements Anthropic has with leading hyperscalers.

Our recent datapoints highlight very strong demand for Nvidia’s upcoming Grok 3 LPX inferencing accelerator, while we expect stand-alone Vera to carve an estimated low-teen unit market share within in the AI server CPU market in 2027.

Nvidia remains one of our top large-cap ideas owing to market share leadership and ongoing share gains including in inferencing, recent reacceleration in revenue growth driven by agentic AI expected to gain further momentum next year, along with strong initial traction for incremental-revenue architectures including stand-alone Vera and Groq 3 inference accelerator.”

This research note is reproduced verbatim from the issuing firm. Price Target never edits, paraphrases or alters analysts’ words — we only republish them in one place.