The picks and shovels of the AI revolution
Las mismas empresas que protagonizaron el auge tecnológico de los años noventa —Dell, Cisco, Intel, Nokia, Micron— vuelven a ocupar el centro del escenario bursátil, no por nostalgia, sino porque la inteligencia artificial necesita una infraestructura física que ellas llevan décadas fabricando. En un ciclo que desafía la memoria colectiva del colapso del año 2000, los supervivientes de aquella era se revalorizan a velocidades que evocan, con inquietante precisión, los mismos excesos que entonces los hundieron. La pregunta que subyace no es técnica sino histórica: ¿puede la demanda real sostener lo que el entusiasmo ha construido tan deprisa?
- Dell ha subido más de un 200% en seis meses y Micron roza el 1.000% en un año, cifras que no se veían desde los días previos al estallido de la burbuja puntocom.
- No son las nuevas estrellas de la inteligencia artificial las que lideran el rally, sino las empresas veteranas que fabrican los servidores, chips y redes sobre los que la IA funciona.
- Nokia, que llegó a perder el 98% de su valor tras el colapso de 2000, ha encontrado una segunda vida como proveedora de la columna vertebral física que los centros de datos de IA exigen.
- Los analistas advierten que la verticalidad de las subidas y la sensación de que las reglas antiguas ya no aplican reproducen con exactitud las condiciones que precedieron al desplome de hace veintiséis años.
- La clave del debate no es si la demanda de infraestructura existe —existe y es masiva— sino si los mercados están descontando un futuro lo suficientemente duradero como para justificar estas valoraciones.
Algo inesperado está ocurriendo en los mercados: las grandes tecnológicas de los años noventa —Dell, Cisco, Intel, Nokia, Texas Instruments, Micron— están protagonizando una de las remontadas bursátiles más llamativas de la última década. No son las nuevas plataformas de inteligencia artificial las que encabezan el rally, sino los supervivientes del colapso del año 2000, aquellas empresas que pasaron décadas reconstruyéndose en silencio.
Los números son difíciles de ignorar. Dell ha subido más de un 200% en seis meses. Micron ha cruzado el umbral del billón de dólares de capitalización tras dispararse cerca de un 1.000% en un año. Cisco ha superado los máximos que registró en el año 2000. Intel acumula casi un 450% de revalorización en doce meses. Nokia, que llegó a perder el 98% de su valor tras el estallido de la burbuja puntocom, ha encontrado una nueva razón de ser en la conectividad que la inteligencia artificial demanda.
El motor de todo esto no es la especulación sobre la IA en sí misma, sino la infraestructura que la hace posible. Cada centro de datos necesita servidores, chips de memoria, equipos de red y conexiones de alta velocidad. Son las palas y los picos de esta nueva fiebre del oro digital, como las describe el analista Manuel Pinto de XTB. Estas empresas no están apostando por el futuro de la IA: están fabricando el presente físico sin el cual ese futuro no existe.
Sin embargo, la velocidad y la verticalidad de las subidas despiertan una memoria incómoda. Los analistas señalan que el patrón recuerda demasiado a los meses previos al colapso de 2000: revalorizaciones abruptas, sensación de que las métricas tradicionales han quedado obsoletas, y un mercado que parece descontar un crecimiento sin límite. La pregunta que nadie puede responder con certeza es si esta vez la demanda es lo suficientemente real y duradera como para sostener lo que el entusiasmo ha construido en tan poco tiempo.
Fifty years after the dot-com bubble collapsed and wiped fortunes from the market, something unexpected is happening. The same companies that led the technology boom of the 1990s—Dell, Cisco, Intel, Nokia, Texas Instruments, Micron—are surging again, this time riding the wave of artificial intelligence infrastructure demand. They are not the startups or revolutionary new players capturing the AI boom. They are the old guard, the survivors, and they are moving faster than anyone anticipated.
Dell's stock tells the story most dramatically. After two decades of recovery from the 2000 crash, the computer maker's shares have climbed 100 percent in a single month, 180 percent in three months, and more than 200 percent over six months. The velocity is striking enough to recall the frenzy that preceded the bubble's burst. Cisco, the networking company, follows a similar arc though less sharply angled—up 31 percent in the last month, 60 percent in three months, after years of stagnation. Intel has accelerated to a 130 percent gain in three months and nearly 450 percent over the past year. Texas Instruments has climbed 40 percent in three months and 75 percent in six. Micron Technology, the semiconductor maker, has become one of the rally's brightest stars, crossing the trillion-dollar market capitalization threshold with a 150 percent jump in three months, 300 percent in six months, and nearly 1,000 percent over the past year.
Even European technology companies from that era are participating. Nokia, the Finnish giant that lost nearly 98 percent of its value after the dot-com implosion, has found new life. The company is benefiting from the infrastructure buildout required by artificial intelligence—the networks, the connections, the physical backbone that AI systems demand. Manuel Pinto, an analyst at XTB, frames the phenomenon with a question: Are the 1990s back in fashion? The data suggest at least a partial answer.
What is driving this resurgence is not sentiment or nostalgia. It is infrastructure. Every new data center requires servers. Every data center needs networking equipment, storage systems, memory chips, industrial components, and high-speed connections. The companies that manufacture these things—the picks and shovels of the AI revolution, as Pinto describes it—are the ones capturing value. Dell is experiencing record demand for AI servers. Cisco has surpassed the highs it set during the year 2000, powered by its exposure to networks and data centers. These are not speculative bets on artificial intelligence itself. These are bets on the physical machinery required to make AI work at scale.
The parallel to the pre-2000 environment is not lost on analysts. The vertical climbs, the speed of revaluation, the sense that old rules no longer apply—these echo the conditions that preceded the bubble's rupture. Pinto and others are asking whether these companies can sustain their gains or whether the market is pricing in a future that will correct as abruptly as it did twenty-six years ago. The question hanging over the rally is whether history will repeat itself, or whether this time the infrastructure demand is real enough, durable enough, to hold the gains.
Citas Notables
One of the most surprising phenomena this year is that some of the biggest winners in artificial intelligence are not new companies or revolutionary startups, but firms that already led the technology bubble of the late 1990s.— Manuel Pinto, analyst at XTB
Every new data center needs servers, networking equipment, storage systems, chips, industrial components, and high-speed connections. The AI revolution is benefiting not just those building the models, but those supplying the infrastructure.— Manuel Pinto, analyst at XTB
La Conversación del Hearth Otra perspectiva de la historia
Why are 1990s companies suddenly winning in an AI boom? Shouldn't new companies be the ones capturing value?
Because AI infrastructure isn't built by software alone. Someone has to manufacture the servers, the chips, the networking equipment. The old companies already know how to do that at scale.
But these stocks are moving so fast—Dell up 200 percent in six months, Micron up 1,000 percent in a year. Doesn't that feel like a bubble forming?
It does feel familiar. The velocity is striking. But there's a difference: in the 1990s, people were buying companies with no revenue. Now they're buying companies with proven manufacturing capacity and actual orders for equipment.
Nokia lost 98 percent of its value after 2000. How does a company come back from that?
By becoming essential to something new. Nokia makes networking infrastructure. When AI requires massive connectivity, Nokia becomes relevant again. The company didn't change much—the world's needs did.
So the risk isn't that these companies are overvalued, but that demand could disappear?
Exactly. If AI infrastructure buildout slows, or if the market decides it's overbuilt, these stocks could fall as fast as they rose. The physical infrastructure boom is real, but booms can end.