All hope for this bull market seemed to rest on Nvidia�s third quarter earnings. What would normally have been an event confined to the technology sector has turned into a global macro catalyst, with one company now treated as a bellwether for the entire market. Late October earnings had already shown signs of life, Alphabet beat expectations and its shares rallied strongly in the days that followed, helped by major announcements around its flagship artificial intelligence model, Gemini. In the space of a few quarters it has moved from being perceived as a laggard to a genuine contender on several artificial intelligence benchmarks. Most large cap reports looked solid on paper, the kind of numbers that in a normal year would have been followed by quietly positive guidance into year end, however this is not a normal year.
Meta provided the first obvious sign of that, despite posting a strong quarter, with sales up more than twenty per cent and significant ongoing investment in artificial intelligence infrastructure, the stock was punished. After earnings it suffered a sharp double digit drawdown from its October high before finally stabilising again. Many other companies with meaningful exposure to artificial intelligence, whether partial or total, experienced similar turbulence. Yet one key player still had not shown its hand and the market was clearly waiting for Nvidia before deciding what to do with the broader artificial intelligence trade.
For some time Nvidia has been treated as almost untouchable, the apparent Midas touch it has shown in any sector it enters is something markets have rarely seen since before 2019. The 19th of November felt like a very long time away for anyone watching the artificial intelligence theme, and the wait only amplified the sense of dread some analysts were projecting onto the sector and, by extension, onto global markets. When Nvidia finally reported, it once again delivered outstanding year on year sales growth of around sixty per cent and maintained extraordinary margins. Yet the market reaction was notably muted. Historically a print like that would have triggered a raucous rally, with Nvidia and its peers enjoying double digit gains, this time it was little more than a brief pop. The share price rose by roughly five per cent before fading again as sceptics resurfaced with talk of an artificial intelligence bubble and warnings of an overdue correction.
The equity price has ballooned over the past year, far beyond most early projections, and in that sort of environment it is easy for investors to forget that risk management and diversification are non negotiable, the hot hand fallacy has crept back in. Valuation and positioning data still point towards Nvidia eventually retracing this recent pullback, but over a longer horizon it is difficult to ignore the psychological damage caused by constant bubble narratives. That fatigue is starting to show. Nvidia�s elevation to saviour of technology and perceived driver of the wider bull market only increases the pressure on the next earnings report. If these third quarter numbers were not enough to satisfy the market, what will the fourth quarter have to look like? It is a vicious cycle that naturally deepens investor anxiety. By the time the next quarter arrives this may all look like textbook fear mongering. The odds are still stacked in favour of the artificial intelligence bull market, but we have clearly reached a phase where sentiment and positioning can override even the strongest performance indicators if the crowd chooses a particular story, rational or not.
While this drama has played out in equities, commodities have quietly staged a bullish turn. Gold, and more recently silver and copper, have all moved higher, with copper in particular hitting headlines on the back of significant supply concerns. Gold, once seen primarily as a steady store of value, has been propelled into a position of leadership. A softer dollar and growing expectations that the Federal Reserve could cut rates in December have driven a surge in investor and central bank demand. On a month on month basis the moves look modest, but year to date the metal has moved far beyond most earlier forecasts. On paper that might appear to be a simple positive sign yet in reality it has the feel of a classic late cycle signal. Gold is often used as a hedge against inflation and broader macro stress, and strong, persistent rallies have historically aligned with troubled backdrops. The surge that began in the late nineteen nineties and extended into the aftermath of the financial crisis is a good example.
Silver has followed a similar pattern, trading close to twenty per cent higher than last month and printing all time highs around fifty five dollars an ounce. It shares the same macro drivers as gold, but its thinner market and higher volatility add an extra layer of tension. Together, gold and silver add to the evidence that the physical market is expressing more caution than headline equity indices suggest. Copper, meanwhile, links the artificial intelligence story back to the real economy. Data centres, electric vehicles, grid upgrades and renewable projects are all copper intensive. Tight supply and underinvestment have helped push prices to elevated levels, which means the same artificial intelligence buildout that drives Nvidia�s revenue also shows up in demand for copper and other industrial metals.
Against this backdrop I have been pushing ahead with EireQuant alongside my university work. This week I finished the first draft of the EQ Studio MVP frontend. It is a locally hosted desktop application that brings the backend model creation and backtesting engine into a no code environment. The goal is to let non technical users learn and test their own strategies against decades of historical data. The recent challenge has been adapting a backend that is wired around command line flags into something a graphical frontend can orchestrate cleanly, that requires fresh database and application interfaces and a more detailed engineering view where model configuration is customisable down to a very fine level. It has forced me to recreate whole modules and temporarily pause work on the current Sentinel model, which uses a newly engineered proportional integral derivative controller that treats equity gains and losses as inputs into a constantly adapting baseline. In backtests this has capped maximum drawdown at roughly five per cent, compared with fifteen to twenty five per cent for the stronger EQ Core models.
Alongside Studio I have been expanding the content of the bidaily reports, which I plan to make public around the new year. These reports will include daily profit and loss from pipeline decisions, relevant macro and equity headlines with sentiment analysis, brief updates on project progress, and a monthly deep dive case study on a topic I feel strongly about. On the more immediate side I have shipped the data files for the EQ Core Mark 1 model, which will be wired into the site this week. It is the latest evolution of the EQ Experimental model, with tighter controls and a re engineered model controller that delivers stronger metrics and return profiles.
The run in to the end of the fourth quarter will be interesting, important questions remain for the technology giants that are not matching Nvidia�s pace and for the Federal Reserve as it decides whether to validate the hedging we are seeing in gold or hold firm and allow markets to work off this anxiety on their own. In either case investors and analysts have to stay anchored to the core tenets of responsible investing, diversification, risk awareness and intellectual honesty. Past performance is not protection, and analysis has to be grounded in broad market evidence rather than simple bullish optimism.