Musk's "Objectively #1" Grok 4.6 Claim Lands as MSFT Falls 2.25%

Elon Musk unveiled Grok 4.6 and declared it "Objectively #1" in intelligence, speed and cost, and Microsoft Corp (NASDAQ: MSFT) traded at $492.48, down 2.25% on the day, as of 18:45 GMT on 12 August 2026.

Elon Musk has put a new model into the field and given it the bluntest possible marketing line. Grok 4.6, unveiled Wednesday, was declared by Musk to be "Objectively #1" in artificial intelligence — specifically on intelligence, speed and cost, the three axes that matter most to anyone deciding where to route a production workload.

Whether that claim survives independent testing is a separate question from how the market reacted to it. And on Wednesday the market reacted in a very specific direction: away from Microsoft Corp (NASDAQ: MSFT), which traded at $492.48 as of 18:45 GMT, down 2.25% from the previous close of $503.81. The stock spent the session in a range of $491.52 to $501.50, meaning it was sitting near the low end of its own day when the quote was taken.

A red tape in a green market

The context is what makes the Microsoft move interesting rather than routine. The broad tape was up. The S&P 500 proxy SPY stood at $773.55, ahead 0.39%. The Dow 30 proxy DIA was at $537.99, up 0.13%. And the Nasdaq 100 proxy QQQ — the index most loaded with the companies that live and die on AI expectations — was the strongest of the three at $725.35, up 0.96% and trading inside a $722.92 to $727.25 band.

So the tech complex was bid and Microsoft was sold. That is the shape of a company-specific story, not a sector story. On a day when the AI-heavy index gained, Microsoft underperformed it by a wide margin.

The 24/7 Wall St report that carried the news framed the timing as awkward for two companies Microsoft has committed billions to — Anthropic and OpenAI. You can read the original write-up at 24/7 Wall St. The logic is straightforward even if the causation is not provable from a single session's tape: if a rival model genuinely wins on all three of intelligence, speed and cost, then the value of a large equity and compute position in a frontier lab is worth less than it was the day before.

Why "cost" is the word that should worry incumbents

Model leaderboards change hands often enough that a claim to be first on raw intelligence is, on its own, close to a monthly event in this industry. Speed matters more, because latency determines what kinds of products you can build. But cost is the claim with the sharpest commercial edge.

Enterprise buyers do not pay for benchmark scores. They pay per token, per query, per seat. A model that is merely competitive on quality but materially cheaper to run will take share from a better model that costs more, because most real-world tasks — summarisation, classification, extraction, code completion — do not need the frontier. If Grok 4.6 is genuinely the cheapest way to get frontier-adjacent output, the pressure lands on pricing across the whole field, and pricing pressure at the model layer flows straight into the economics of the labs that sell access.

That is the mechanism by which a claim from one company can dent the perceived value of investments in two others. It is not that Anthropic or OpenAI loses customers overnight. It is that the price they can charge, and therefore the path to profitability that justifies multibillion-dollar outside capital, gets harder to defend.

What Microsoft actually has at stake

Here the honest answer is that the public record in this lead does not give a dollar figure. Microsoft's relationship with OpenAI is the most heavily scrutinised corporate partnership in software, and the lead confirms the company is betting billions across the two labs named. But no specific investment total, ownership percentage or revenue-share term is established by the facts available, and inventing one would be worse than leaving the gap visible.

What can be stated is the market's arithmetic on the day. A 2.25% single-session decline in a company of Microsoft's scale is a large absolute number, and it happened while the rest of large-cap technology rose. Investors were repricing something specific.

Three things are worth separating:

• Equity exposure. Stakes in private AI labs are carried at values that depend on the labs' own forward pricing power. A cheaper competitor compresses that.
• Compute revenue. Cloud providers earn on the training and inference their partners run. That revenue is comparatively insulated — models still need chips, whoever wins.
• Product moat. Microsoft's own assistant and developer tools are built on partner models. If a rival model is cheaper and faster, the competitive question is whether Microsoft's distribution is worth more than the model underneath it.

The third point is the one that historically has protected Microsoft. Distribution into corporate IT has beaten better technology many times. But it is also the point that a genuine cost advantage attacks most directly, because procurement teams can now compare line items.

The claim still needs verification

"Objectively #1" is a marketing phrase, not a measurement. The word "objectively" is doing a lot of work in a field where benchmark selection, prompt engineering and evaluation harnesses all shift results, and where every lab publishes the comparison set that flatters it. Independent evaluation over the coming weeks — third-party leaderboards, developer throughput tests, real per-token pricing rather than promotional pricing — is what will determine whether this is a genuine repricing event or a one-day headline.

Also worth noting: the lead attributes the unveiling to SpaceX rather than to Musk's AI venture. Musk's corporate structures have been reorganised repeatedly, and the attribution of a model release across his companies is itself a detail investors will want clarified, because it affects which entity captures the revenue and which shareholders benefit.

What to watch from here

The immediate tell is whether Microsoft's underperformance persists into subsequent sessions or reverses. A single day inside a $491.52 to $501.50 range, on a day the Nasdaq 100 proxy gained 0.96%, is a signal but not a trend. If the gap between Microsoft and QQQ keeps widening, the market is treating the model-cost story as structural.

Second, watch for published pricing. A stated per-token price that undercuts the field is checkable in a way that a benchmark chart is not, and it is the number that will move enterprise procurement.

Third, watch the language in Microsoft's own disclosures about how it carries and describes its AI partnerships. Valuation marks on private stakes are where a competitor's price war eventually shows up in reported numbers — with a lag, and usually after the share price has already moved.

Frequently asked questions

What did Elon Musk claim about Grok 4.6?

Musk declared Grok 4.6 "Objectively #1" in artificial intelligence, specifically citing intelligence, speed and cost. The model was unveiled on 12 August 2026. The claim has not been independently verified, and benchmark leadership in AI is contested and frequently changes hands between competing labs.

How did Microsoft shares react?

Microsoft Corp (NASDAQ: MSFT) traded at $492.48 as of 18:45 GMT on 12 August 2026, down 2.25% from the previous close of $503.81. The stock ranged between $491.52 and $501.50 during the session, sitting near the low end of that band when the quote was taken.

Was the wider market down that day too?

No. The S&P 500 proxy SPY rose 0.39% to $773.55, the Dow 30 proxy DIA rose 0.13% to $537.99, and the Nasdaq 100 proxy QQQ gained 0.96% to $725.35. Microsoft fell while the AI-heavy technology index rose, making the move company-specific rather than sector-wide.

Why would a rival AI model hurt Microsoft?

Microsoft has committed billions to Anthropic and OpenAI. If a competing model is genuinely cheaper and faster at comparable quality, it compresses the pricing power those labs have, which in turn makes the value of large outside investments in them harder to justify at prior marks.

Why does cost matter more than benchmark scores?

Enterprise buyers pay per token, per query or per seat, not for leaderboard positions. Most production tasks — summarisation, classification, extraction, code completion — do not require frontier capability, so a cheaper model that is merely competitive on quality can take substantial share from a better but pricier one.

How much has Microsoft invested in OpenAI and Anthropic?

The available reporting confirms Microsoft is betting billions across the two labs but does not establish a specific investment total, ownership percentage or revenue-share term. Investors looking to size the exposure precisely will need Microsoft's own disclosures on how it carries and values those private stakes.

Sources

• SpaceX Just Unveiled Grok 4.6. Musk Calls it “Objectively #1” In AI. Anthropic, OpenAI and This Stock May Be in Trouble. — 24/7 Wall St

Photo: AlphaTradeZone · Pexels Licence — source