On Thursday, AI applications exploded, and concepts such as media, film and television, publishing, and games collectively soared.

As of midday trading, Liujin Technology has risen more than 16%, and Rongxin Culture has once risen and stopped in the intraday period, but now it is over 11%. Many stocks such as Cultural Investment Holdings, Beijing Culture, Longpan Media, and Gravity Media have risen and stopped, and Guomai Culture, Aofei Entertainment, Kunlun World Wide, and Tom Cat have followed suit.

It is worth noting that Longfan Media, which has continued to “run wild” recently, has been rising and halting for 5 consecutive trading days. Currently, it is reported at 25.18 yuan, with a total market value of 11.19 billion yuan.

In response to this, Dragon Media issued a clarifying announcement last night, stating that it has recently noticed that some media and investors have classified the company as a “data element” concept stock and understood the “multi-dimensional frontier knowledge service product database” as a data element concept. butThe company does not belong to the concept of data elements, and content resources are far from reaching the “massive” level.

Tech giant “Big Killer” debuts

Overnight, Google and AMD went big and new, releasing blockbuster products with the “strongest” AI direction.

On Wednesday local time,AMD held a “Promising AI” press conference and made a high-profile announcementThe new MI300 series AI chip,Includes the MI300A and MI300X chips.

Among them,MI300X seriesIt is a large GPU with the memory bandwidth required for leading generative AI and the training and inference performance required for large language models;MI300A seriesThe integrated CPU+GPU, based on the latest CDNA 3 architecture and Zen 4 CPU, can provide breakthrough performance for HPC and AI workloads.

Currently, the MI300A has entered mass production, and the MI300X has begun shipping.

Regarding AMD's “new Wang Bang,” Meta and Microsoft said they will buy AMD's new artificial intelligence chip as a replacement for Nvidia. Additionally, Oracle's cloud computing will also use this type of chip.

AMD CEO Lisa Su expects the AI chip market to reach $400 billion or more by 2027, and she hopes AMD has a sizable share of this market.

If AMD's latest high-end chip, the MI300X, starts shipping early next year, the performance is good enough and the price is cheap enough for technology companies and cloud service providers that build and service artificial intelligence models, then it may reduce the cost of developing artificial intelligence models.

In addition to AMD's release of the “strongest computing power,” Google released the “strongest AI model.”

On Wednesday local time, Google announced the launch of the artificial intelligence model Gemini.

According to the benchmark results given by Google, Gemini showed “the most advanced performance” in many tests, even completely beating OpenAI's GPT-4 in most benchmarks.

Google has released three different versions for different scenarios:Gemini Ultra, Gemini Pro, and Gemini Nano.

Gemini Ultra - The strongest and largest model for handling highly complex tasks.
Gemini Pro - the best model for expanding various tasks.
Gemini Nano - the most efficient model for devices such as phones.

As for the native multi-modal Gemini, Google says this is so farThe largest, most capable, and most versatile large language modelIt surpassed GPT-4 in various fields such as text, video, and voice.

Also, starting December 13, developers and enterprise customers will be able to access Gemini Pro's Gemini API through Google AI Studio or Google Cloud Vertex AI.

AI applications are in full bloom

Currently, the AI track is in full swing.

In addition to overseas tech giants continuously updating major products from computing power to big models, the AI products of some “rookie” startups are also “rising”, and AI applications are beginning to blossom.

Some time ago,The popularity of Pika, an AI video generator tool, is out of the ring.It also confirms that native AI applications are exploding at an accelerated pace.

According to reports, Pika has raised 55 million US dollars through three rounds of financing. Currently, the valuation of Pika Labs has reached between US$200 million and US$300 million.

Open Source Securities points out that AI multi-modality, especially AI video generation capabilities, is developing rapidly, or further improve applicability in the gaming field. Recently, the AI multi-modal model has continued to achieve breakthrough development.

The upgrading of the multi-modal capabilities of the AI model and the reduction of deployment costs will further enhance its application scope and applicability. As a commercial application of content forms such as text, images, sound, and video, games are expected to make better use of the multi-modal model on the R&D side, improve content production efficiency, and increase content richness. At the same time, some AI-based NPC designs may further enhance user experience and willingness to pay, and open up room for growth in overall game revenue.

CITIC Securities said that the official release of the multi-mode Gemini model can, on the one hand, expand application scenarios, and on the other hand, bring about a continuous upgrade in computing power requirements. The agency said,We continue to be optimistic about the future of the AI industry, believing that the subsequent release of models such as GPT-5 will also bring more catalysts.

In the short term, the release of Gemini will further bring more expectations for the multi-modal model. For the industry, multi-modal materials will drive an increase in computing power demand; in the medium to long term, it is expected that the multi-modal model upgrade will enrich the usage scenarios of related products and add cost optimization brought about by hardware upgrades and algorithm optimization. The progress of 2C products is worth looking forward to. We continue to be optimistic about the long-term impact and changes of this wave of generative AI on the technology industry, and continue to focus on leading manufacturers in computing power, algorithms, data, and applications.